The Secret Behind The Tortoise and the Hare: Information Design in Contests

Abstract

I analyze the optimal information disclosure problem under commitment of a “contest designer” in a class of binary action contests with incomplete information about the abilities of the players. If the contest designer wants to incentivize the players to play in equilibrium a particular strategy profile, she can design an information disclosure rule, formally a stochastic communication mechanism, to which she will commit and then use to “talk” with the players. The main tool to carry out the analysis is the concept of Bayes Correlated Equilibrium recently introduced in the literature. I find that the optimal information disclosure rules involve private information revelation (manipulation), which is also cost-effective for the designer. Furthermore, the optimal disclosure rule involves asymmetric and in most cases correlated signals that convey only partial information about the abilities of the players.